Every insurer is investing in AI to sharpen underwriting, speed up pricing, and reduce loss ratios. Yet, most initiatives underdeliver. Not because the models are wrong, but because the data feeding them is fragmented, inconsistent, or simply too slow to reach the point of decision. The result is promising pilots that never scale, and models that should not be trusted with real risk decisions.
In this session Dr Marco Rauch, Partner at d-fine's Insurance Practice, will make the case that the real underwriting edge isn't the model, but rather the data foundation beneath it. Using examples from our experience with building modern data platforms with Databricks at European insurance companies, we'll share what separates AI initiatives that deliver measurable underwriting and pricing impact from those that stall in pilot purgatory. We'll cover:
- Why data architecture, not model sophistication, is usually the binding constraint on AI value in insurance
- What "AI-ready" data looks like in an underwriting context and how far most insurers still are from it
- How leading insurers are sequencing investment: building the foundation first to unlock faster, more reliable AI models second
- A short outlook on where advanced GenAI models or agentic underwriting assistants/Genie can meaningfully move the needle once that foundation is in place
After the presentation there will be plenty of time to stay for networking and enjoying another coffee. Dr Julian Leppin, our Expert in Solvency II and AI-enhanced Risk Management, who spoke at the event in May, will be joining as well and would be happy to answer any follow-up questions.
We at d-fine Stockholm are happy to invite you to our second Insurance Breakfast with our d-fine experts where we combine a conversation about building AI-ready data foundations with coffee, breakfast, and a relaxed start to the day.
Speaker
Dr Marco Rauch, Partner, Insurance Practice
d-fine
